Towards a Handwritten Text Interpretation Framework for Ancient Spanish Manuscripts

EDUARDO XAMENA; CARLOS ISMAEL OROZCO; CARRASCO CABRERA GASTÓN

Salta

Simposio; 1er Simposio Argentino de Imágenes y Video. 48 JAIIO; 2020

DI-UNSa

Abstract. Handwritten Text Recognition is a widely studied researchtopic. We implement novel binarization methods in order to preprocesshandwritten text images efficiently and accurately, acquiring adequatebinary black-white images for later recognition processes. Afterwards, thecharacters present in the documents are used to train and evaluate deep-learning mechanisms for the recognition task. Our framework providesgood source images for the recognition phase in terms of noise removaland processing of low contrast images. Besides, the process of characterrecognition is also improved by means of deep-learning techniques.